Research On Fault Diagnosis Method Of Fuzzy Petri Network With Distributed Power Distribution Network

In power system distribution network,the rapid and reliable diagnosis and recovery of fault is a hot spot for many scholars.The fault diagnosis method of distribution network used to use only one kind of fault information,switch quantity information or electrical quantity information,no combination of two kinds of information is used.In addition,with the development of clean energy,the distribution of distributed generation is becoming a trend.In addition,with the development of clean energy,Distributed Generation has become a trend to connect to the power grid.More and more distributed power sources have made a great impact on the relay protection and network topology.In the power failure information system,there is a wealth of logical and temporal relationship between electrical quantities and switching quantities.Integrating the inherent relationship information into fault diagnosis will make the analysis more rapid and reliable.Fuzzy Petri net is a kind of human cognitive reasoning method has been applied to various engineering fields,also play a role in the direction of fault diagnosis.But the traditional method of fuzzy Petri nets,there are still some problems,such as:the initial values of fuzzy Petri nets by artificial experience to give,but the statistical data in the past is still relatively scarce,experience of protection and circuit breaker for lack of credibility in accordance with fault tolerance;the model parameters are more subjective and therefore not reasonable.;not adaptable to changes in the network structure.Therefore,this thesis which improving fuzzy Petri net model,use BP algorithm parameters on the weights and belief for training.Meanwhile making full use of distribution network in the electrical and timing information to achieve a more reliable and accurate fault diagnosis of distribution network.This thesis mainly do the following work:(1)Two information systems for fault diagnosis of distribution network are simply studied.Combined with the type and flow of fault information,the information obtained from the failure of power grid is analyzed.The influence of distributed generation on relay protection and fault diagnosis of distribution network,especially the relationship between DG capacity and protection setting value,is analyzed and simulated,and the characteristics of various power grid components are analyzed.(2)The calculation method of fuzzy Petri net initial base reliability is improved to solve the fault information exists misoperation,refused to move the electric quantity and switch quantity logic relationship,troubleshooting information is invalid,while addressing the protection and circuit breaker initial library credibility rely on expert experience problems.We improve the diagnosis model of fuzzy Petri nets,and use the BP algorithm to learn the parameters of fuzzy Petri nets,get rid of the dependence on subjective factors,and make the diagnosis more convincing.(3)From the angle of fault current transient,the fault area of distribution network containing DG is analyzed.According to the result of current wavelet transform at fault protection operation,the fault and non fault area are compared,providing basis for correcting fault information and eliminating the influence caused by DG.This thesis improves the fuzzy Petri net model of the radiated distribution network containing DG and analyzes its adaptability.The feasibility of the method is verified by an example analysis.